KR910015962A - Voice signal processing device - Google Patents

Abstract

No content

Description

Voice signal processing device

Since this is an open matter, no full text was included.

FIG. 2 is a block diagram showing an embodiment of the audio signal apparatus according to the first invention of the present invention, FIG. 3 is a graph showing the cepstrum peak in the embodiment, and FIG. 4 is a second diagram of the present invention. A block diagram showing an embodiment of an audio signal processing apparatus according to the invention.

Claims (6)

A frequency analysis means for frequency analysis of the audio input signal, a pitch extraction analysis means for pitch extraction analysis of the output of the frequency analysis means, a pitch rum extraction means for detecting a pitch in the pitch extraction analysis output, and the pitch extraction analysis Based on the average value calculating means for calculating the average value level in the analysis output of the means, the pitch detection information of the pitch detecting means and the average value information of the average value calculating means, the vowel is determined based on the pitch, and the average value information And a vowel / consonant determination means for determining vowels and consonants based on the consonants. Band dividing means for band-segmenting an audio input signal, thrip analysis means for performing a spectral analysis of the band division output, and a spectral peak at the spectral analysis output of the spectral analysis means; On the basis of the peak detection means, the average value calculation means for calculating an average level in the chop strm analysis output of the shock strut analysis means, the peak detection information of the peak detection means, and the average value information of the average value calculation means. And vowel / consonant determination means for determining vowel based on the peak, and for determining vowel and consonant based on the level of the average value information. The vowel / consonant determination means according to claim 2, wherein the vowel / consonant determination means comprises: a first comparator for comparing the detected peak in the peak detection means and a threshold set by the threshold setting portion, and a predetermined average set by the average value calculation means and the threshold setting portion; And a vowel / consonant determination circuit for determining the vowel and the consonant and outputting the result based on the comparison results of the first and second comparators. . Frequency analysis means for frequency analysis of the audio input signal, cepstrum analysis means for performing the spectral analysis of the frequency analysis output of the frequency analysis means, and the cepstrum in the spectral analysis output of the cepstrum analysis means. A peak detection means for detecting peaks, an average value calculating means for calculating an average value level in the chop strum output of the chop strum analyzing means, peak detection information of the peak detecting means and average value information of the average value calculating means. Based on the peak, the vowel is determined, the vowel / consonant determination means for determining the vowel and the consonant based on the average value information and the level, and the determination result in the vowel / consonant determination means. Erasing coefficient setting means for setting an erasing coefficient by using a noise predicting means for predicting a noise component by applying a voice signal converted to the puuri; Noise prediction output of the noise predicting means; an erasing means for inputting an audio signal and an erasing coefficient signal set by the erasing coefficient setting means, and canceling means for canceling a noise component in consideration of the erasing rate from the audio signal; And a signal synthesizing means for synthesizing the signal. Band dividing means for dividing the voice input signal rental division; Chop strut analysis means for performing a spectral analysis on the band split output of the band dividing means; and a spectrum in the spectral analysis output of the spectral analysis means. Peak detecting means for detecting peaks, an average value calculating means for calculating an average value level in the chop stratum analysis output of the shock analyzing means, peak detection information with the peak detecting means, and an average value of the average value calculating means. Based on the information, the vowel is determined based on the peak, the vowel / consonant determination means for determining the vowel and the consonant based on the level of the average value information, and the judgment in the vowel / consonant determination means. An erase coefficient setting means for setting an erase coefficient using the result, a noise predicting means for inputting a voice signal converted into the puuri and predicting the noise component thereof, A side output, the audio signal, and an erase coefficient signal set by the erase coefficient setting means are input to the noise of the tone predicting means, and erase means for canceling a noise component in consideration of the erase rate from the voice signal; And a band synthesizing means for band synthesizing the erase output. 6. The apparatus according to claim 5, wherein the vowel / consonant determination means includes at least a detection peak in the peak detection means and a threshold value set by the threshold setting unit, a first average comparator, a predetermined average value set by the average value calculation means, and a threshold setting unit. And a vowel / consonant determination circuit for determining a vowel and a consonant and outputting a result based on a comparison result between the first and second comparators. ※ Note: The disclosure is based on the initial application.